Further Research Into Medicare Advantage Upcoding, Documentation Accuracy Needed

In the final segment of this interview with Andrew S. Oseran, MD, MBA, MSc, advanced heart failure and transplant cardiologist at Beth Israel Deaconess Medical Center, he suggests future research areas and policy changes to address concerns about Medicare Advantage (MA) payment accuracy.

Watch parts 1 and 2 for insights into the key objectives and findings of his study, "Prevalence of Chronic Medical Conditions Among Medicare Advantage and Traditional Medicare Beneficiaries."

This transcript has been lightly edited for clarity; captions were auto-generated.

Transcript

Given concerns about potential "upcoding" in MA plans, what further research is needed to determine whether coding differences reflect more comprehensive documentation or inappropriate inflation of risk scores?

That's a really important question, I think it's a point worth emphasizing. If we take our study results at face value, and we assume that MA and fee-for-service populations are indeed similar in terms of their medical complexity, then the higher risk scores that are observed could be due to 1 of 2 things.

On the one hand, it could be that MA plans are just more fully capturing their beneficiaries' comorbid conditions but that they're doing so accurately. Alternatively, it could be that these risk scores represent inappropriate upcoding of conditions.

I think future research to tease that apart, fundamentally, will require, one, a gold standard way to ascertain what conditions a beneficiary does or does not have, and then researchers would need to be able to look at claims data for that same patient to see what comorbidities are actually coded.

If you can find a way to do that on a large scale, at a population level, then I think you can start to understand to what extent risk scores submitted by MA plans are accurate and to what extent they represent upcoding or inappropriate coding. We're definitely thinking about ways to do that in future studies, using similar linked data sets to the ones that we used in the study we're talking about today.

Based on your findings, what policy changes, if any, should CMS consider to ensure accurate and fair payment structures for MA plans?

A lot of what we found informs policy debates that are already taking place, that are ongoing. I think our findings add to a growing body of evidence that the risk adjustment methodology used to calculate MA plans may need to be reformed. If the MA population really is no more medically complex than the fee-for-service population, these plans should not be receiving more funds from the federal government unless the care quality and the outcomes that their beneficiaries obtain are far superior, and it's just not clear that that's the case from the available evidence.

Policy makers continue to discuss potential reforms to this system. These reforms range from things like across-the-board reductions in MA risk scores to correct for more intensive coding, which already takes place but could be increased. It could include things like adding new elements to risk scores, data derived from pharmaceutical usage or patient surveys, things that may be less prone to gaming, to more fundamental changes to the way that we reimburse MA plans.

I think, ultimately, we'll need to invest in studying all of these solutions and probably more of them.